Abstract
A method for spatio-temporally smooth and consistent estimation of cardiac motion from MR cine sequences is proposed. Myocardial motion is estimated within a four-dimensional (4D) registration framework, in which all three-dimensional (3D) images obtained at different cardiac phases are simultaneously registered. This facilitates spatio-temporally consistent estimation of motion as opposed to other registration-based algorithms which estimate the motion by sequentially registering one frame to another. To facilitate image matching, an attribute vector (AV) is constructed for each point in the image, and is intended to serve as a "morphological signature" of that point. The AV includes intensity, boundary, and geometric moment invariants (GMIs). Hierarchical registration of two image sequences is achieved by using the most distinctive points for initial registration of two sequences and gradually adding less-distinctive points to refine the registration. Experimental results on real data demonstrate good performance of the proposed method for cardiac image registration and motion estimation. The motion estimation is validated via comparisons with motion estimates obtained from MR images with myocardial tagging.
| Original language | English |
|---|---|
| Pages (from-to) | 2514-2526 |
| Number of pages | 13 |
| Journal | Pattern Recognition |
| Volume | 42 |
| Issue number | 11 |
| DOIs | |
| Publication status | Published - 2009 Nov |
Bibliographical note
Funding Information:A 4D deformable registration method for estimation of cardiac motion from MR image sequences was presented, and experimentally tested. The cardiac motion estimation was formulated as a 4D image registration problem, which simultaneously considers all images of different time-points and further constrains the spatiotemporal smoothness of estimated motion fields concurrently with the image registration procedure. Also, compared to other motion estimation methods that use relatively simple features such as curvature of the left ventricular border, our method uses a rich set of attributes, including GMIs, to distinguish the corresponding points across different time-points, thereby maximally reducing ambiguity in image matching. Finally, by selecting the active points hierarchically based on the degree of distinctiveness of their attribute vectors, the proposed registration algorithm has the potential to produce a global solution for motion estimation. The performance of this 4D registration method for cardiac applications has been evaluated by visual inspection as well as quantitative validation using tagged MR images. Experimental results demonstrate good performance of our method in estimating cardiac motion from cine MR sequences. The motion estimates are statistically similar to the motion of tag lines obtained from tagged MR images. In addition, the radial and circumferential strain estimates obtained by our methods from cine MR images compare favorably with those obtained from co-registered tagged MR images. Cine MR images are commonly acquired in a clinical setting, and the use of our algorithm may obviate the need to acquire additional tagged images to estimate myocardial motion. In addition, our method allows for dense estimates of myocardial motion, which is not restricted to the intersection of tag lines, thus helping detect abnormal myocardial motion and also potentially improving early diagnosis and treatment planning of cardiomyopathies. About the Author —DINGGANG SHEN received all of his degrees from Shanghai JiaoTong University. He is currently an associate professor in UNC-Chapel Hill. Before moving to UNC-Chapel Hill, he was a tenure-track assistant professor in University of Pennsylvania (UPenn) and a faculty member in Johns Hopkins University. Dr. Shen is on the Editorial Board of Pattern Recognition , International Journal of Image and Graphics , and International Journal for Computation Vision and Biomechanics . He also served as a reviewer for numerous international journals and conferences, as well as NIH, NSF and other grant foundations. He has published over 200 articles in journals and proceedings of international conferences. His research interests include medical image analysis, pattern recognition, and computer vision.
Keywords
- Cardiac motion estimation
- Image registration
- Spatio-temporal normalization
ASJC Scopus subject areas
- Software
- Signal Processing
- Computer Vision and Pattern Recognition
- Artificial Intelligence